between the fem and the do... (when manual calculations do not quake with the fem.. 'help!!)

[Fulvio Romano]

the material in the fem is hyperelastic, linear and isotropic if you impose it as such, otherwise you can also realize materials that are not.
for a steel, however, seems to me a very acceptable hypothesis.

right, as long as you can then interpret the numerals near the carvings.

other note: but who said that to the courses of science of construction and in those of construction of machines is not said that the attainment of the yield at a point does not involve the collapse of the structure? and that the hypothesis of Saint-venant are seen as perfectly responsive to reality?

university of napoli federico ii, prof. luciano nunziante.
For the record, I passed the full grade exam, and then I had to figure out how the materials work in the subsequent exams.
the professor in question, in fact, in the introduction to the course explained well how things are, but then during the course it was easily allowed to take the hand from cauchy!

the saying that there are no bad pupils, but only bad teachers, unfortunately it is not always true.

We have an example in the forum. at least by reading the various posts, but are not subject to discussion :tongue:

 

[Zaxxon]

Bye to all,
I would like to refocus on the subject fem.
the problem in my opinion is not so much the university notions dictated by professors (they can also be failable, but studying for textbooks you can not fail to understand certain notions such as plasticity, carving etc... in juvinall are repeated endlessly and I do not conhoscate engineer who has never read it).
the problem lies in the poor knowledge of the fem in general, no one today studies first how the fem works and then model to seen ragion...the more settle to make 4 tutorials put there and then claim to do analysts. but if you ask each of them what the tension is at one point and if they are sure the stutter comes.. .
the tragedy is that at the base of an acceptable result there is mesh, if you do not waste time is useless....the results will always be doubtful. with free meshators unfortunately the vase of pandora has been opened and together with it chaos in the preventive since discretization weighs 50% of the work.

returning to the "round" treatment in general, I personally prefer to work on a simplified model and then use the submodelling technique if necessary.

Hi.

♪

 

[Ing.Vedder]

thanks to all for the ton of answers, all excellent ideas of reflection.

I try to answer something, but I'm sorry, I'm home now and I'm weird. .

1- the piece is in simple s235jr, for which I thought to an admissible of 160mpa

2- Let's admit that while shaping the edge with a proper curvature radius the problem persists... is it the case to worry?! (I say we admit why I am not so interested in the single case but rather to learn how to make me an idea of what I am analyzing and what the fem is telling me. . )

3) I am not interested in fatigue, this team is a sort of hook that takes a load and translates it for about 300 mm. will make 100/200 cycles and then the piece will no longer serve

4) place here a photo with the measurements of the ear..if you should check it "hand" how would you proceed? !

I would like to see the tension in the section I indicated with the section line a-a then. .

wx = bh/62 = 6000 mm3
a-a = 8500*83 = 705500 n*mm
= 705500/6000 = 117.6 n/mm2 <-- ok

This stupid stupid stupid stupid calculation of me.. But he doesn't find a match with the one found by the fem... Perhaps I should add a carving factor as indicated by the good Roman lightning. .

Boh.

thanks again to all for the thousand ideas and reflections. .

ing.vedder

 

[folle76]

Well, it certainly lacks the factor of carving (although I do not know if it is a case contemplated by peterson (I think not), you certainly do not find it if you leave the living edge). the remaining tensile stress will make only 1 mpa.
if for your case are worth all assumptions about the possibility of neglecting the peak tension that have been exposed in this discussion then consider team checked from the static point of view.
I do not deal with printed materials, but geometries, such as saying, elusive and evasive. that is why I am forced to calculate average tensions on the sections and compare them with those of yielding, but above all with those of pieces already in exercise for years, which, deo gratias, have never come down. I don't have any idealized beams to cling to or calculate carving factors to which I am peacefully shaking.
let's say that more than designing game to discover the differences between the tension of a new piece and one, morphologically similar, that has been standing for years before I was born and before the saint-venant discovered the beam.

 

[Fulvio Romano]

the moment at the a-a section should be the same.
In my opinion, it is the case of:
- add traction too, little stuff, but why not put it on us. then use the sigma of von mises
- I'd pass the cut, indeed, it should be neglected.
- by the cutting factor, in the first line, multiplies by 3 the maximum sigma. It's a wrong account, but you should get incredibly close to reality
- Most importantly. How does the load look like? is loaded with extreme calm, or by a "black" operator? In that case, the impulsive load breaks the piece because it is higher than the static...how to evaluate it? Good question...
When the load is moved, are we sure that the "donder" does not strive to twist the piece? is it not the case of putting some welded ribbing that increases torsion stiffness?

 

[Zaxxon]

I think there's no more information,
the fem can be used to calculate the cutting coefficient of any geometry. Now, if the edge is at 90° there is no verse, the more you put the mesh the more the value of the tension rises (until the infinite in theory), but if that edge is connected with the radius "right" or presumed such, the voltage values can be reasonably correct and, evaluating the stress at a reasonable distance from the edge, by relation can calculate the intaglio factor.

♪

 

[cacciatorino]

more inflate mesh plus the value of tension salt (until in theory)

I miss the physical motivation of this statement, it also applies when modeling the material with its yield curve, or only if you consider the infinitely elastic linear material? In this second case maybe this can be understandable, but modeling is conceptually wrong.

 

[Fulvio Romano]

I miss the physical motivation of this statement, it also applies when modeling the material with its yield curve, or only if you consider the infinitely elastic linear material? In this second case maybe this can be understandable, but modeling is conceptually wrong.

I believe that motivation is not physical, but mathematics.
If you think of hydrodynamic analogy, then the "flow " of tensions must have a discontinuity. in reality the mesh rounds the discontinuity as much as it is crude. Inflating the mesh you get an increasingly accentuated discontinuity.
in the real world the discontinuity does not exist.

 

[gerod]

in fact the model with mesh is mathematical, discreet made of discrete numbers.
to simulate reality, such elements should tend to the point, which is unlikely.
In my opinion, I tend to continuity, i tend to reality.

to explain, we should think about the mathematical model of the fems. I read a great book on the subject (I don't remember the title and I don't have it under hand) made by an Italian engineer expert at fem. treating is quite simple and explains very well the concept of mesh, of accumulation of tensions and especially when a result is good.

 

[Zaxxon]

both gerod and lightning are right,
so if you want to make an estimate of the tension you need to model the round, perhaps in a later step using a submodel as I had already anticipated.
Then we did not talk about the problem of the shape of the elements of the mesh....for the reasons mentioned above a "point" element (tetrahedri, prisms etc.) of nature leads to have singularities in the calculation of the derived quantities (type stress, but also a thermal flow, a convection coefficient etc...) and therefore is contraindicated when the quantities that you want to measure are not primary for the element...
so when showing you important results always take a look at mesh before colored mappines! :

See you soon!

♪

 

[caste81]

with the live edge inflating mesh efforts tend to infinitely for a matter of fracture mechanics.
the only way to avoid this is to collect the edge.
the team in question once connected can be considered a hook and therefore it should be calculated as a solid at large curvature.

Hi.

 

[AMinati]

The asme rules div. 2 and the European correspondents en 13445.3 (related to the calculation of pressure vessels) contain specific parts for the analysis carried out by finite elements. tensions are grouped into categories according to their characteristics and for each of these categories different limits are established which are respectively 1.0, 1.5 or 3.0 times the permissible voltage. They also establish methods to measure maximum values near discontinuity. in this way the peak values reported by the programs due to the affect effects are resized. Of course the effects of carving will still be taken into account in the case of particular fatigue subjects for the known implications and risks.

as the last beach... the resistance reserves of materials are many..., but never rely on them!


hoping for help
aminati


p.s. some software performs automatically or almost this type of analysis (linearization of tensions).

 

[Matteo]

according to me (and in this is the difference between the analyst fem and the designer who adapts to make calculation) the mathematical result of the model fem must always be binding and in favour of safety. I explain: if a live edge appears beyond the permissible load (e.g. 1/2 of the breaking sigma) then the piece does not pass the verification (the discourse for which "at exepreience" can be tolerated punctual peaks up to 2 or 3 times the admissible... but what does "punctual" mean?? do you refer to the size of the red zone?? absurd. we forget that the results plot is a "mediate" representation of the nodal or centriidal loads. true: there is the risk of being too much in favor of security, but the only thing to do at this point is to improve the accuracy of the calculation to approach it to reality, clear that the game becomes more complex and less easy to manage for those who are not specialist). For example, we can drop the simplistic hypothesis of always linear material (studying local plasticization), or we can detail more geometry (collections), or still simulate the presence of mechanical treatments that through the creation of "residue" tensions (compressions, "stiratures" that make the material transversally isotropic) reduce the effect of fatigue in critical points, or take into account the speed of application of the load etc.